/visual

name: visual
description: Generate on-brand visuals and infographics for LinkedIn posts and newsletters using Nano Banana MCP (Gemini AI image generation). Final step of the content pipeline. Follows a guided flow: analyze the content, suggest visualizable concepts, propose visualization approaches, then generate on-brand visuals. USE THIS SKILL WHEN user says "create a visual", "make a visual", "generate an infographic", or wants a graphic to go with a post or newsletter.
disable-model-invocation: true

Infographic Generator

Create professional infographics in the BenAI brand style for LinkedIn and newsletters. Uses Gemini AI via the Nano Banana MCP server for image generation.

**Connector:** Nano Banana MCP (Gemini AI image generation).

**Core rule:** Quality is set by the thinking BEFORE the prompt. Spend 70% of effort on what to visualize, 30% on how.

**UX rule:** When the AI can make a good default, make it silently and let the user override. Never ask the user to confirm something the AI already decided.

Files load lazily, one phase at a time. Do NOT load reference files up front. See the Reference File Load Map at the bottom.

---

When This Skill Loads

1. Run Phase 0 (silent config check). 2. If the API key is NOT configured: jump to Phase 5-ALT (API Key Setup). Do NOT ask about content. 3. If the API key IS configured: begin Phase 1.

---

UNDERSTAND

Phase 0: Silent Config Check

Run silently. No questions yet.

mkdir -p .infographic/images .infographic/prompts

if [ -n "$GEMINI_API_KEY" ]; then
  echo "API key found in environment"
elif [ -f .env ] && grep -q "^GEMINI_API_KEY=.\+" .env; then
  echo "API key found in .env file"
else
  echo "API key not configured"
fi

• Key found (env var OR .env): proceed to Phase 1.
• Key NOT found: jump to Phase 5-ALT.

Then check for brand config (first found wins): `./branding.md`, then `.infographic/brand.md`. If found, `cat` it and acknowledge briefly ("Found your brand context, ready to create."). `branding.md` is written by the `/ads-creative` skill (colors, fonts, voice, audience); use it for on-brand palettes, typography, and style anchors.

---

Phase 1: Content Intake

No reference files. No AskUserQuestion. Say:

> "What content should I turn into an infographic? Paste text, share a URL, or point me to a file."

Acknowledge receipt briefly, then go to Phase 2.

---

Phase 2: Analysis + First Choices

>>> READ references/content-analysis.md NOW

2.1: Four-Layer Analysis

Apply the four-layer framework in `content-analysis.md`. Present the analysis labeled by layer (Narrative, Themes, Claims, Information) with honest strength ratings (Weak / Moderate / Strong).

2.2: Auto-Decide Output Mode

Auto-decide the output mode from the layer strengths using the decision table in `content-analysis.md` (Auto-Intent Detection + Hybrid Sub-Modes). State the decision inline with the analysis, including the hybrid sub-mode when relevant. Do NOT ask the user, unless both Layer 3 and Layer 4 are weak (the only case where you ask for guidance). The user can override any auto-decision in free text.

2.3: Batch Platform + Type

ONE `AskUserQuestion` with TWO questions. Always ask BOTH regardless of the auto-detected mode, so the user always has the choice.

questions:
  - question: "Where will this be posted?"
    header: "Platform"
    options:
      - label: "LinkedIn (Recommended)"
        description: "4:5 portrait, optimized for feed engagement"
      - label: "Instagram"
        description: "1:1 square or 4:5 portrait"
      - label: "Twitter/X"
        description: "16:9 landscape for timeline"
      - label: "Presentation"
        description: "16:9 landscape for slides"

  - question: "What type of visual should this be?"
    header: "Type"
    options:
      - label: "Infographic (Recommended)"
        description: "Clean data visual using BenAI brand system. Cards, scorecards, diagrams. Best for frameworks and comparisons."
      - label: "Editorial illustration"
        description: "Artistic, cinematic scene. Moebius linework, surrealism, or graphic novel style. Best for philosophical claims."
      - label: "Risograph / analog print"
        description: "Paper grain, muted palette, slight imperfections. Tactile and warm. Best for personal, human content."
      - label: "Minimalist conceptual"
        description: "Single powerful image. New Yorker cover energy. Best for one strong metaphor."

Move the "(Recommended)" tag to the editorial option if the auto-detected mode is Editorial Illustration or Editorial-led Hybrid.

The type selection OVERRIDES the auto-detected mode. Apply the Type Selection Override table in `content-analysis.md`, then state the locked mode.

---

PLAN

Phase 3: Concept Selection

Extract at least 5 concepts. Always extract from BOTH Layer 3 and Layer 4 (per the extraction rules in `content-analysis.md`), so the user can pick a concept that shifts the output type. Label every concept with its layer and what it produces.

Present all 5+ with `AskUserQuestion`. Use the actual claims or data shapes as labels, not topic names; include the layer tag in the description.

question: "Which of these should we visualize?"
header: "Concept"
options:
  - label: "[First few words of data shape 1]"
    description: "Layer 4 (infographic). [Why it is a strong visual candidate]"
  - label: "[First few words of claim 1]"
    description: "Layer 3 (editorial). [Why it is a strong visual candidate]"
  [... never fewer than 5 options]

Ordering: concepts matching the locked mode come first (Layer 4 first for Information Graphic / Data-led Hybrid; Layer 3 first for Editorial), but always include both layers. If the user picks a concept from a different layer than the locked mode, silently switch the mode to match and state the switch.

Never proceed without a user selection.

---

Phase 4: Visualization Approach

>>> READ references/visualizat